On Cell觸控感應器產品混線生產製程之參數優化—彩色濾光片廠為例

本技術報告探討運用彩色濾光片(color filter，簡稱CF)製程的生產線作業模式，來混線生產On Cell觸控感應器的問題，針對機台設備提出參數設定之優化方法，並以個案G公司作案例說明。面對On Cell觸控製程導入與CF產品做混線生產，其過往的CF生產線作業流程經驗無法套用，需要發展一套新型混線生產及參數優化模式。G公司先前以試誤法做參數優化，無法有效達到縮短生產前置時間及產品不良率管控目的，因此有必要提出一套科學且可提高良率的方法。針對此問題，研究方法分成三大步驟：（一）統計客訴反應的觸控產品不良狀況並選出預計改善的影響因素（只針對某較常用的產品規格）、（二）找出造成觸控產品不良狀況的原因、相關製程及重點機台設備（只針對某較常用的產品規格）及（三）利用實驗設計方法為重點機台設備的重要製程參數進行參數值的優化（只針對某較常用的產品規格）。其中第二步驟較複雜，所使用的方法又包含：(i)特性要因圖：目的是列出生產線（人、機、料、法）中可能影響的因素；(ii)品質跨職能流程圖：目的在檢視生產流程是否有重複或者可造成製程不良的生產程序；(iii) 品質機能展開：目的是對預計改善的不良狀況，由大到小逐一縮小範圍並且加以收斂至生產線可能造成不良狀況的機台設備、參數等因素；(iv)因果矩陣：目的是經收斂後的柏拉圖比例關係，羅列影響製程的關鍵因子；(v)失效模式分析(FMEA)：目的在對影響製程的關鍵因子做細部收斂分析求出RPN值，觀察其分數相對較高的項目，就可視其為重點機台設備、參數。依據上述所提研究方法，發現蝕刻段製程的參數為重點改善參數，包含蝕刻機台的噴灑壓力、基板的搬送速度及蝕刻液溫度等三項，發現噴灑壓力這因子是顯著因子，而搬送速度及溫度兩因子不顯著，再者考量蝕刻液存放時間可較長且不增加加熱成本的因素，另外縮短週期時間的考量下將搬送速度設定機台最高值，可增加稼動率的產能提升，最後提出這三個因子的參數值組合。針對本技術報告提出的參數值組合，G公司進行小量生產測試，從G公司生產型號GP5A3054的On Cell產品檢驗結果來進行改善成效分析，包含兩部分(1)不良率分析及(2) 及R管制圖分析。從不良率分析發現改善前不良率為27.27%，樣本數為55，95%信賴區間為 (0.1550 , 0.3903)，改善後為不良率2%，樣本數為48，不良率的95%信賴區間為=(0.0196 , 0.0596)平均而言不良率降低25.27 %。從及R管制圖分析可看出，改善前有較多的點超出管制界限，而且超出甚多表示製程中心位置偏移規格較遠，改善後的及R管制圖幾乎都在管制界限內，表示製程穩定且製程中心接近規格值。接著再製造成本降低部分，因為不良率的降低使得不良品成本降低，而成本降低的金額會隨著稼動率改變而不同，從稼動率為100 %、80 %、60 %分別來看，相對可節省的每日成本分別11,200,000 NT、8,960,000 NT、6,720,000 NT

關鍵字

彩色濾光片；On Cell觸控感應器產品；薄膜電晶體液晶顯示器製造廠；混線生產；波賽頓On Cell觸控感應器產品

並列摘要

This technical report explores the use of color filters (abbreviated as CF) process line operating mode to mixed line production On Cell touch sensor problem, an optimization method for parameter setting of machine equipment, and in the case G Company for case study. On Cell touch sensor the control program to import products made with CF mixed line production, its past experience in CF line processes can’t be applied to optimize the model requires the development of a new line of production and mixing parameters. G Company previously made trial and error method parameter optimization, can’t effectively achieve shorter production lead times and product defect rate control purposes, it is neces-sary to propose a set of scientific methods and can improve the yield. To solve this problem, research methods into three steps: (a) the reaction of customer complaint statistics touch products and selected factors adverse conditions are expected to improve (more commonly used only for a certain product specifications), (b) identify the cause Touch optimize product adverse conditions reasons related process and focus on machine equipment (only for a more common product specifications) and experimental de-sign (c) the use of focused machine equipment of important process parameters parameter value (only for certain over the Common product specifications). Where the second step is more complex, but also contains the method used: (i) fish-bone diagram: The purpose is to list the factors of production lines (human, machine, material, method) that might affect; (ii) the quality of cross-functional flowchart: The pur-pose of viewing the production process has to be repeated or whether the process can cause poor production procedures; (iii) Quality function deployment: the purpose is expected to improve the poor conditions, descending one by one narrow and converge to the production line may be adverse conditions machine equipment, parameters, and other factors; (iv) the causal matrix: The aim is by Plato proportional relationship after convergence, a list of key factors affecting processes; (v) Failure Mode Analysis (FMEA): The purpose of affecting the process of the key factors to do RPN values obtained detailed analysis of convergence observed a relatively high fraction of the project, we can focus on machine equipment, parameters. According to the above mentioned methods, found that the parameters of etching ma-chine is focused on improving the parameters, including etching machine of the spray pressure, the conveyor speed and the etchant temperature, found spray pressure which fac-tors are significant factors, while conveying speed and the temperature of the two factors was not significant, moreover considering etchant can be stored for longer without in-creasing the cost of heating elements, in addition to shortening the cycle time of the con-veyance under the consideration of the highest speed setting machine, increase the utiliza-tion rate of production can be improved, finally This combination of parameter values proposed three factors. For combinations of parameter values presented in this technical report, G companies small production test models produced from G's On Cell GP5A3054 product testing results to improve effectiveness analysis consists of two parts (1) Product defect rate analysis and (2) and R control chart analysis. analysis the Product defect ratio, that not improved yield before was 27.27% , sample size was 55 confidence interval (0.1550, 0.3903), the Product defect ratio improved to 2% for the 48 samples, 95% confidence interval for the Product defect ratio = (0.0196, 0.0596) lower average the Product defect ratio 25.27%. From the analysis and R control charts can be seen, there are more points before the improvement beyond the control limits, but many said the process beyond the center position offset spec-ifications far, the improved and R control charts in almost all control limits indicates the process is stable and close to the center of the process specification values. Then again re-duce manufacturing costs partly because lowering rates reduce the cost of making defective products and the amount of cost reduction will change as the utilization rate is different from, utilization rate of 100%, 80%, 60%, respectively, the relative availability Daily cost savings were 11,200,000 NT, 8,960,000 NT, 6,720,000 NT